Get the Agent OS + join 4,000+ founders inside the AI Profit Boardroom → Join AIPB ($59/mo)
Game

Flightsim

Flight Simulator — take off, fly over terrain, full flight HUD, land on the runway.

CategoryGame
Models tested25
Scored20/25
Avg score7.50/10
WinnerFusion

What I asked each model — the Flightsim prompt

Every model on this page got this exact prompt inside the Agent Operating System: Flight Simulator — take off, fly over terrain, full flight HUD, land on the runway.

Single HTML file out. No iteration. No examples in the system prompt. Whatever each model produced on the first run is what's on this page. 25 frontier models have attempted it so far: Claude Fable 5, Fugu Ultra 1.1, Fusion, Gemini 3.6 Flash, GLM-5.2, GPT-5.6 Sol, Grok, Grok 4.6, Hy3, Inkling, Kimi K2.7, Kimi K3, MiniMax M3, Hermes MoA, Muse Spark 1.2, Opus 4.8, Claude Opus 5, Qwen 3.8, Qwen 3.7, Claude Sonnet 5, DeepSeek V4 Pro, DeepSeek V4 Flash, Kimi K2.7 · Fast, Kimi K2.7 · No-Think, Kimi K2.7 · Quality.

Why this task matters. Flightsim is a textbook test of game-class capability — the kind of build that exposes whether a model is doing pattern-matching or actual reasoning. A model that ships this in one shot is usually safe to wire into your agent loop for harder tasks of the same shape.

How each model handled Flightsim

Ranked by my 0–10 score from the source comparison guides on agentos.guide. Click any to play the actual one-shot HTML the model produced.

Claude Fable 5 Anthropic
• 8.0/10

What I saw: One-shot: real aircraft model (wings/prop/gear) on runway w/ centerline + edge lights + tower, complete glass PFD (speed/alt tapes, pitch ladder, heading ribbon), autopilot demo rolls. Alt tape shows negative rows. Eyeballed.

▶ Play Claude Fable 5's attempt →
Fugu Ultra 1.1 Sakana AI
🥉 8.4/10

What I saw: Polished flight sim with a detailed aircraft model, full HUD (airspeed/alt/VS/heading tape/attitude indicator), runway with markings, hangar, control tower and terrain — plus a combat layer with visible drones and cannon reticle. Very strong and shippable, but the enemy at this frame is just a single distant drone and it sits right at the top of the field's best rather than clearly beating it.

▶ Play Fugu Ultra 1.1's attempt →
Fusion OpenRouter
🥇 8.7/10 · showcase · game-director

What I saw: SHOWCASE BUILD (threejs-game-director): livery aircraft w/ animated control surfaces + prop blur, real flight model (stall/ground effect/graded landings), canvas runway + PAPI + windsock airport, glass-cockpit strip (IAS/ALT tapes, horizon, heading), auto-demo. Eyeball-gate passed.

▶ Play Fusion's attempt →
• 8.1/10

What I saw: Strong, cohesive flightsim: clean low-poly terrain with mountains/trees, full HUD (airspeed, altitude, VS, heading tape, artificial horizon, throttle), a visible aircraft on the runway, plus a combat/landing loop. No enemies visible in this frame and it's a mid-runway static shot, so it lands just short of the best but is clearly shippable.

▶ Play Gemini 3.6 Flash's attempt →
GLM-5.2 Zhipu / Z.ai
• 8.0/10

What I saw: 66KB · plays clean · three, webgl

▶ Play GLM-5.2's attempt →
GPT-5.6 Sol OpenAI
🥉 8.4/10

What I saw: Renders cleanly with a polished, cohesive HUD—airspeed/altitude tapes, compass, throttle, nav map, brackets and flight-path marker—and a believable runway-perspective terrain with a chase-cam aircraft, hitting all brief elements (takeoff, terrain, HUD, landing assist). Loses a touch versus the top field on visual richness (flat terrain, faint aircraft model) but it's strong and shippable.

▶ Play GPT-5.6 Sol's attempt →
Grok xAI
• 8.0/10

What I saw: 9KB · plays clean · three, webgl, input, rAF

▶ Play Grok's attempt →
Grok 4.6 xAI
• 5.0/10

What I saw: One-shot score. The world is the best in the run, a low-poly valley with a river, hangar, control tower, marked runway and a sunset sky, plus a proper attitude ball and compass tape. The flight model was dead though: after nine seconds of throttle and five of back-pressure the HUD still read IAS 042KTS, ALT 0000FT, VS +000FPM, unchanged at every sample. Grok 4.6 fixed its own build from that evidence and the live demo now rotates and climbs to 992ft, though it still stalls if you over-pitch.

▶ Play Grok 4.6's attempt →
Hy3 Tencent Hunyuan
• 6.8/10

What I saw: Clean HUD with working attitude indicator, heading tape, throttle/gear/score panels, and a decently modeled aircraft with wings, nav lights and prop; but the terrain reads as an empty green haze with no visible runway, trees or structures from this altitude, leaving the world flat and generic compared to the field's best.

▶ Play Hy3's attempt →
Inkling Thinking Machines
• 3.5/10

What I saw: HUD panel is polished and on-brand, but the screenshot shows a broken render — a giant close-up brown cross (the plane fills the frame) against a black void with no visible terrain, runway, sky, or clouds; the camera/scale is clearly wrong and the sim reads as non-functional.

▶ Play Inkling's attempt →

The winner on Flightsim

Fusion took gold on this task. showcase · game-director.

What I saw: SHOWCASE BUILD (threejs-game-director): livery aircraft w/ animated control surfaces + prop blur, real flight model (stall/ground effect/graded landings), canvas runway + PAPI + windsock airport, glass-cockpit strip (IAS/ALT tapes, horizon, heading), auto-demo. Eyeball-gate passed.

See Fusion's full model card: /models/fusion. Direct head-to-head against the runner-up: Fusion vs Claude Opus 5.

Every attempt — live, playable

Side by side. Click any tile to run that model's actual one-shot HTML in a new tab.

How I scored Flightsim — methodology

Three axes, 0–10 each, averaged. Runs: drop the .html in a browser; if it opens to a broken page, it scores zero. Hits the brief: did the model ship the thing the prompt asked for, or a different thing it found easier. Looks good: visual polish, motion, interactivity — where most of the gap between gold and silver lives.

My scores trace back to the source comparison guides on agentos.guide. See the full methodology page for data provenance, including which source guide each cell's score came from.

The same stack Julian uses

Run this stack yourself.

Every demo on this bench was built inside the Agent Operating System — one prompt, one shot, single HTML file out. The Agent OS, the prompts, the templates, the weekly walkthroughs and 4,000+ founders shipping with it every day all live inside the AI Profit Boardroom.

4,000+founders
258documented wins
38countries
$59/momonthly